What is Ethereum? — The Daily Coins

Ethereum is the second-largest cryptocurrency by market cap, but calling it a cryptocurrency understates what it is. Bitcoin is a payment network. Ethereum is a programmable platform that happens to have a payment network as one of many features. If Bitcoin is digital gold, Ethereum is closer to a global, public computer that anyone can deploy code to and that no single entity controls.

This guide is the long version. We cover the mental model, how it actually works, the major upgrades that have shaped what Ethereum is in 2026, the Layer 2 ecosystem, staking, the DeFi and NFT applications built on top, and the real risks. Read it top to bottom if you are new. Skip to the sections that interest you if you already own ETH.

A 90-second mental model

Picture a single global computer that anyone can run code on, anyone can read the results of, and no single entity controls. That computer is the Ethereum Virtual Machine (EVM). The fuel for the EVM is a token called Ether (ticker ETH). Every operation on the EVM — sending money, executing a contract, minting an NFT, calling a function on another contract — costs a small amount of ETH, called gas. Validators run the EVM, propose blocks, and earn ETH for keeping it honest.

The simplest way to picture it: imagine a worldwide spreadsheet where every cell is also a tiny program. The state of every account, every token balance, every smart contract’s internal memory — all of it lives in that shared spreadsheet. Anyone who runs the Ethereum software sees the same state. Any change to the state has to be requested by a transaction signed by the account that has the right to make it. The network rejects invalid transactions and applies valid ones in the same order on every node.

That is what “world computer” means in practice. It is not fast. It is not cheap. But it is credibly neutral and unstoppable in a way no centralised database can be. The applications that have grown on top of it — decentralised exchanges, lending protocols, NFT marketplaces, stablecoins — are interesting because they inherit those properties.

How it works

Underneath the analogy, Ethereum is concrete software. A few key pieces are worth understanding.

Smart contracts. A smart contract is code deployed to Ethereum that runs exactly as written. It can hold funds, transfer them based on rules, interact with other contracts, and emit events that the outside world can listen to. Once deployed, the rules cannot be changed unilaterally — there is no admin override unless the developer explicitly built one in (and many have, with mixed results). Most contracts are written in Solidity, a JavaScript-flavoured language designed for the EVM. A growing share use Vyper or other languages that compile to EVM bytecode.

Gas. Every EVM operation has a cost measured in gas units. Sending ETH is cheap (21,000 gas). Calling a complex DeFi contract can cost hundreds of thousands of gas. The user pays a gas price per unit, denominated in gwei (one billionth of an ETH). Total transaction cost = gas used × gas price. When the network is congested, gas prices spike — which is why an NFT mint can cost $5 one day and $300 the next.

Validators. Since the Merge (September 2022), Ethereum has used proof-of-stake. Anyone with 32 ETH can become a validator. Validators take turns proposing new blocks and attesting to blocks proposed by others. Honest behaviour earns rewards. Dishonest behaviour — proposing two conflicting blocks, going offline for long stretches, or trying to game the consensus rules — gets the validator’s stake slashed. Slashing is the cryptoeconomic enforcement that keeps the network honest. As of 2026, the network has well over a million active validators.

Finality. Once a block is included and two more epochs (about 12.8 minutes) pass without challenge, the block is considered final. Reversing a final block would require an attacker to deliberately burn enormous amounts of staked ETH. This is a much stronger guarantee than Bitcoin’s probabilistic finality, where you wait six confirmations and call it good enough.

The Merge

In September 2022, Ethereum executed the most ambitious live upgrade in crypto history. It switched from proof-of-work (the energy-intensive mining mechanism Bitcoin still uses) to proof-of-stake. The transition was completed in a single block. No price chart moved meaningfully, no funds were lost, and energy use dropped by roughly 99.95% overnight. It was, by software standards, one of the most boring fireworks shows ever staged.

The Merge changed several things. Energy use went from comparable to a small country down to roughly that of a few thousand US households. Issuance dropped sharply — under proof-of-work, the network was paying out roughly 13,000 ETH a day in block rewards; under proof-of-stake, it pays out closer to 2,000 ETH a day to validators. Combined with the fee burn introduced earlier (more on that next), ETH became net deflationary during periods of high network activity.

The Merge also moved Ethereum decisively into territory where staking yield matters. ETH became, for the first time, an asset that produced cash flow. The implications for valuation models are still being worked out by the industry.

EIP-1559 and fee burning

One step before the Merge, in August 2021, Ethereum activated EIP-1559. This was a fee market overhaul that fundamentally changed how transaction pricing works. Instead of users blindly bidding gas prices, each block has a base fee that adjusts up or down based on whether the previous block was over or under target capacity. Users can add a small priority fee (tip) to incentivise inclusion.

The critical part: the base fee is burned. Permanently destroyed. Removed from supply. This means every transaction reduces the total ETH that will ever exist. Over time, if the burn outpaces new issuance to validators, ETH supply shrinks. During periods of high activity in 2023 to 2024, this happened on a sustained basis. The phrase the community uses is “ultrasound money” — a riff on Bitcoin’s “sound money” thesis. It is meant a bit tongue-in-cheek, but the underlying mechanic is real.

The combination of EIP-1559 plus proof-of-stake is what makes ETH an unusual asset: it produces yield for stakers, it burns supply when used, and the burn scales with demand. That is a tighter feedback loop than most assets have.

The Layer 2 ecosystem

Ethereum main net (Layer 1) is intentionally slow and expensive. The design goal is to keep the global state small enough that anyone with a consumer laptop can run a node and verify the chain themselves. Scaling, in Ethereum’s roadmap, does not happen by making L1 faster. It happens by building Layer 2 networks on top.

An L2 batches thousands of transactions, executes them off-chain, and posts the compressed result back to L1. Users pay a small fraction of L1 fees but inherit L1’s security: if anyone tries to cheat, the proof system on L1 catches them. The major L2s in 2026 are:

Arbitrum One. The largest L2 by total value locked. An optimistic rollup — it assumes transactions are valid by default and only computes proofs if someone challenges. Cheap fees, mature ecosystem, full EVM compatibility.
Optimism / OP Mainnet. The other major optimistic rollup. Pioneered the OP Stack, an open framework other chains use to launch their own rollups (Base is one of them).
Base. Coinbase’s L2, built on the OP Stack. Best onboarding from fiat thanks to Coinbase’s rails. Growing fast.
zkSync Era. A zero-knowledge rollup. ZK rollups use cryptographic proofs that mathematically guarantee correctness — no challenge period needed. More complex to build, but they offer stronger finality and (eventually) better economics.
Polygon zkEVM, Linea, Scroll, Starknet. Other significant ZK rollups, each with different tradeoffs around EVM equivalence, proving system, and ecosystem.

For most everyday Ethereum activity in 2026 — trading on a DEX, lending, NFT mints, gaming — users are on an L2, not L1. L1 is increasingly settling for institutional flows and L2 commitments. This is the “rollup-centric roadmap” the Ethereum Foundation laid out in 2020, playing out in real time.

Staking

You can earn yield on ETH by helping run the network. There are several ways in.

Solo staking. Run your own validator. Requires 32 ETH, technical comfort with a server, and a stable home internet connection. You earn the full yield — roughly 3 to 5% in protocol issuance plus a portion of MEV (more on that below). Slashing risk falls entirely on you, but if you run the standard client software correctly, the risk is small.

Staking as a service. Services like Allnodes or Stakefish run a validator for you. You still need 32 ETH, but they handle the infrastructure for a small fee. You keep custody of the withdrawal key.

Pooled staking. If you do not have 32 ETH, pooling services aggregate smaller amounts. Two patterns dominate. Liquid staking protocols like Lido and Rocket Pool give you a tokenised receipt (stETH, rETH) that represents your staked position and can itself be used in DeFi. Centralised staking on exchanges (Coinbase, Kraken) is the simplest — you click a button and earn yield, but the exchange holds the keys.

Liquid staking is where a large share of ETH ended up. Lido alone has historically held a meaningful percentage of all staked ETH. That concentration is itself a source of debate — see the risks section.

Yields fluctuate. They are higher when fewer ETH are staked and lower when participation is high. In 2026, base protocol yields hover around 3% with another 0.5 to 1% from priority fees and MEV. Real after-tax returns depend heavily on jurisdiction.

DeFi natives on Ethereum

Ethereum’s real value is in the applications built on top of it. A handful are worth knowing, both because they are economically significant and because they explain what the platform is for.

Uniswap. The original automated market maker. Instead of an order book, Uniswap uses liquidity pools — pairs of tokens deposited by users who earn a share of trading fees. Anyone can list a token by creating a pool. This is what made the long tail of token trading possible and what makes Ethereum’s spot DEX flow material to global price discovery for many coins.

Aave. The largest decentralised lending protocol. Users deposit assets to earn yield. Other users borrow against collateral. Interest rates adjust algorithmically based on utilisation. The whole thing runs as a few hundred lines of audited Solidity. No loan officer, no credit check.

MakerDAO (now Sky). The original on-chain stablecoin. Users lock up ETH or other assets as collateral and mint DAI — a dollar-pegged stablecoin overcollateralised by the protocol. DAI has weathered multiple crises (March 2020, Terra collapse in 2022, USDC depeg in 2023) and remained mostly on peg.

Lido. The largest liquid staking protocol. Holds a significant share of total staked ETH. Its stETH token trades in DeFi as a yield-bearing equivalent of ETH.

These four together hold tens of billions of dollars in value. They are the most-tested pieces of public infrastructure in DeFi, and they all live on Ethereum (with deployments on multiple L2s).

NFTs and beyond

An NFT is a token that follows the ERC-721 standard rather than the ERC-20 standard used by fungible tokens. ERC-20 says “these tokens are interchangeable” — your ETH is the same as my ETH. ERC-721 says “each token is unique and ownership is tracked individually.”

This sounds simple but unlocks a lot. Digital art ownership, in-game item rights, ENS names (decentralised name service), real estate deed pilots, ticketing experiments — all of these are NFT use cases. The cultural moment around profile-picture NFTs in 2021 to 2022 distorted the discourse, but the underlying primitive is general purpose and persistent.

NFT activity moved largely to L2s during 2023 to 2024 once gas costs made L1 minting prohibitive for anything except the most premium collections. ERC-1155 (a multi-token standard) and various semi-fungible patterns have also grown. Where NFTs go from here is anyone’s guess — the speculative froth is mostly gone, the technology is intact.

How Ethereum competes

Ethereum is not alone. Solana, Avalanche, Cosmos, Aptos, Sui, and a long tail of newer chains all compete for developer attention and user activity. The competitive landscape is worth understanding because it affects ETH’s long-term thesis.

Solana trades decentralisation for speed. Single-leader sequencer, monolithic chain, much cheaper transactions, lower hardware requirements for users but higher for validators. Has been particularly strong for memecoins, DePIN projects, and applications that need low-latency execution. Ethereum’s response is L2s plus the rollup roadmap.

Cosmos and the app-chain ecosystem trade composability for sovereignty. Each chain has full control over its own stack, including its own validator set and economic policy. Strong for protocols that want maximum customisation; weak for protocols that want easy access to other DeFi liquidity.

Bitcoin layers (Stacks, RGB, Liquid, and emerging zk projects) are starting to build expressive functionality on top of Bitcoin’s settlement layer. Still small but worth watching as a different model for combining Bitcoin’s security with programmability.

Ethereum’s competitive moat is not speed. It is the combination of credible neutrality, a mature security model, the largest pool of audited contracts and tooling, and the most deeply rooted developer ecosystem. Whether that moat holds depends on whether the rollup roadmap delivers cost and speed competitive with monolithic alternatives. The current evidence is mixed-positive: L2s have made Ethereum usable for everyday transactions, but the user experience of bridging between L1 and L2s is still rough.

Real risks

Ethereum has more attack surface than Bitcoin because it does more. The real risks are worth being honest about.

Smart contract bugs. Code is law on Ethereum, which means if the code has a bug, the bug is law. Hundreds of millions of dollars have been lost to vulnerabilities in DeFi contracts — reentrancy, integer overflow, oracle manipulation, signature replay. Reputable protocols audit extensively, but no audit catches everything. If you use a brand-new protocol on day one, you are an unpaid security researcher.

MEV (maximal extractable value). Validators choose which transactions to include in their blocks and in what order. This ordering power has economic value — they can frontrun trades, sandwich users, and arbitrage between DEXs. MEV is technically legal on the protocol; whether it is “fair” depends on who you ask. For users, the practical effect is occasional worse fills and the existence of an entire ecosystem of MEV-aware tools (Flashbots, MEV-Boost) designed to mitigate the worst behaviour.

Validator centralisation. Lido staking, large exchange validators, and a handful of staking-as-a-service providers control significant shares of validator activity. None of this rises to the level of a 51% attack risk in 2026, but the trajectory matters. A more decentralised validator set is healthier for the network than a more concentrated one.

Regulatory stance on staking. The US SEC has, at various points, suggested that staking-as-a-service products may be securities offerings. Different administrations have taken different views. The legal status of yield-bearing ETH products in the largest economy in the world remains an open question, and policy shifts could affect both prices and access.

Complexity risk. Ethereum has more moving parts than Bitcoin. More upgrades, more configurations, more contracts, more L2s, more bridges. Each piece is one more place where bugs can live. The Pectra upgrade and the rollup-centric roadmap are genuine improvements, but they also increase total surface area. This is the tradeoff Ethereum has made: more programmability for more failure modes.

Bridge risk. Cross-chain bridges have been the single largest source of crypto theft in history — multiple incidents over $300 million each. If you bridge assets between L1 and an L2, or between two chains, you are trusting the bridge’s security. Use native bridges where possible; treat third-party bridges with extreme caution.

Pectra and the upgrade roadmap

Ethereum upgrades come in bundles, typically named after either devcons or stars. The major upgrades since the Merge:

Shanghai/Capella (April 2023). Enabled withdrawals of staked ETH for the first time. Before this, validators could stake but could not exit. Shanghai made staking a properly liquid asset class.

Cancun/Deneb (March 2024). Introduced EIP-4844 “proto-danksharding”, which added a new transaction type for L2 rollups to post data much more cheaply. L2 fees collapsed almost immediately, with most rollups going from cents per transaction to fractions of a cent.

Pectra (Prague-Electra). The next major upgrade after Cancun, bundling validator UX improvements, increased maximum stake per validator, and account abstraction primitives. Activated in stages across 2025 into early 2026.

Beyond Pectra, the long-term roadmap continues to be rollup-centric: improving data availability for L2s, hardening the validator set, refining MEV mitigation, and migrating toward statelessness so that nodes do not need to store the entire global state. The cadence is one major upgrade every 12 to 18 months. Compared to Bitcoin, this is breakneck. Compared to most web platforms, it is glacial.

How we cover Ethereum on this site

Ethereum gets daily coverage on The Daily Coins. The relevant pages are:

/coins/ethereum/ — live ETH price, market cap, supply, and our forecasts across all horizons.
/predictions/ethereum/ — the dedicated Ethereum forecast page.
/category/ethereum/ — every Ethereum article we have published.
/category/layer2/ — Layer 2 coverage, focused on Arbitrum, Optimism, Base, and zkSync.

If you only bookmark one, the coin profile at /coins/ethereum/ is the one. It updates continuously and links to the rest.